Dan Foreman-Mackey
Dan Foreman-Mackey
Building on https://github.com/google/jax/pull/21925, this tutorial demonstrates the use of the FFI using `ffi_call` with a simple example. I don't think this should cover all of the most advanced use cases,...
Following up on https://github.com/google/jax/pull/21925, we can update the example code in `docs/cuda_custom_call` to use `ffi_call` instead of manually registering `core.Primitive`s. This removes quite a bit of boilerplate and doesn't require...
I used a combination of ruff's `--fix` option and manual editing to get all the doctrings to the be a little more consistent in formatting. This is a huge PR,...
Would it be possible to give it a transparent background, @lgrcia?
It has come up a few times (most recently in https://github.com/google/jax/issues/23624) that the "vectorized" behavior of `pure_callback` and `ffi_call` is confusing. I'm working on improving that, but in the meantime,...
This PR includes an end-to-end example project which demonstrates the use of the FFI. This complements [the FFI tutorial](https://jax.readthedocs.io/en/latest/ffi.html) by putting all of the code in one place, as well...
These optimizations start to narrow the gap between the limb_dark and starry solution vector computational performance. The main differences: 1. Using the closed form solution for s0 instead of the...
This isn't really meant to be merged, but it's a follow up from https://github.com/wjakob/nanobind/issues/746 demonstrating the test failure with the current master branch of Eigen. This workflow seems to fail...
### Problem description When I build the following simple nanobind module with the current `master` branch of Eigen: ```c++ #include #include namespace nb = nanobind; NB_MODULE(nanobind_example_ext, m) { using Matrix...
The high level change in this PR is to update `ffi_call` to return a callable. Before this change, the `ffi_call` syntax was: ```python ffi_call("target_name", output_type, *input_args, vmap_method="...", **input_kwargs) ``` but,...